Just another Saturday
Alright, two things for today.
Minor RLCS update
The RLCS presentation of November is coming up, and to be honest I am worried about the time limit. I’m trying hard to decide what to remove. While I was on that, I found an error in the Rosetta Stone function for the Iris demo. I then added another simple (not great) function to decode rules.
Not great, but hinting at what can be done, and maybe the future versions of the Rosetta Stone functions could end up being solid. No matter. Good enough for now.
Visuals for ARC AGI 1
I downloaded a couple of the ARC AGI 1 exercises (I need to look into crawling all of them, actually).
But then, I need to visualize the exercises.
So I just put together a couple of functions for that.
Nothing major, but I will say I was having trouble with the image() function, so I’m moving to ggplot. Nicer looking anyway.
Miscellaneous
In the Cybersecurity Market, there are interesting conceptual things to think about. There are things happening in IT, Cybersecurity, etc. that are not perfectly solved. The whole prompt injection/MCP/AI browsers, etc. It’s a mess (from my perspective).
And it’s taking a bit of my thinking time, but as I say: Interesting.
Conclusions
Not much today.
The hardest part of a presentation is apparently to make it as short as possible. And while at it, I should look into making it somewhat entertaining, too.
As for the ARCAGI1, it is still just a nice and interesting thought exercise. Nothing more for now, really, as I have made zero progress towards an actual implementation of anything. In the GEB, D. Hofstadter discusses the Bongard sets and I reviewed the corresponding section of the book this week. It turns out, “my approach” to the project (well, the way I’m thinking I want to approach it) is mostly aligned with that described in the book. But aside from a few notes, yes: Zero progress.
And I need to keep thinking about some stuff.